Project Summary/Abstract Clinical decision support (CDS) is a process for enhancing health-related decisions and actions with pertinent, organized clinical knowledge and patient information to improve healthcare delivery. For example, doctors may have trouble remembering to order all guideline-recommended care for sepsis. CDS delivered via an order set in the electronic health record (EHR) can simplify this process and reduce mortality by bundling the recommended diagnostic and therapeutic orders together. More generally, CDS has the potential to improve patient outcomes by facilitating adherence to evidence-based practices (EBPs). However, it may fail to reach this potential because: (1) the CDS tool is underutilized; (2) the user may not follow the recommended action from the CDS; (3) the recommended action may not lead to the EBP; and/or (4) the EBP may not translate to the expected outcome in a novel population. Healthcare organizations need an efficient, rigorous, and scalable process evaluation method to diagnose when and why CDS is not leading to the intended improvements. Our long-term objective is to empower organizations to efficiently incorporate scientific knowledge into high value clinical care through incremental, data-driven improvements of CDS informed by understanding of the relationships between CDS use, process measures, and patient outcomes. In Aim 1 of this proposal, we will establish the technical feasibility of associating CDS use patterns with process and outcome metrics using EHR-log data through a proof of concept demonstration focused on inpatient treatment of pediatric migraine. Both Children?s Healthcare of Atlanta (CHOA) and Children?s Hospital of Philadelphia (CHOP) have well defined patient cohorts, local clinical guidelines, order sets, and outcome metrics for the care of inpatient pediatric migraine patients. Phrase Health has already developed a commercial CDS analytics product used by 3 institutions that organizes EHR-log data into an intuitive display that provides insight into how all an organization?s alerts and order sets are used. We will leverage these strengths to create a standard database schema for patient cohorts that links to Phrase Health?s existing data model for order set use patterns. We will then verify the accuracy of this association through manual chart review. While we use inpatient migraine as a model, the architecture will generalize across clinical use cases. In Aim 2, we will use user-centered design best practices to develop a visual analytics dashboard for rapid identification of CDS improvement targets. We will then evaluate the effectiveness of the dashboard through scenario-based summative testing, in which we measure how accurately users identify CDS improvement targets with a functional dashboard based on real CDS data extracted in Aim 1. At the end of this project, we will have created and validated a novel, scalable CDS process evaluation tool with demonstrated technical feasibility at two institutions. This will advance healthcare organizations? ability to improve patient outcomes through CDS and prepare us for a Phase 2 application focused on implementation.